E-CAM-S: Physiological Assessment of Delirium Severity: The Electroencephalographic Confusion Assessment Method Severity Score

Abstract

Delirium is a common and frequently underdiagnosed complication in acutely hospitalized patients, and its severity is associated with worse clinical outcomes. The Electroencephalographic Confusion Assessment Method Severity Score (E-CAM-S) is a physiologically based method to quantify delirium severity. This repository provides the data and code used for the following publication: 

The Electroencephalographic Confusion Assessment Method Severity Score (E-CAM-S). van Sleuwen M, Sun H, Eckhardt C, Neelagiri A, Tesh RA, Westmeijer M, Paixao L, Rajan S, Velpula Krishnamurthy P, Sikka P, Leone MJ, Panneerselvam E, Quadri SA, Akeju O, Kimchi EY, Westover MB. Physiological Assessment of Delirium Severity: Crit Care Med. 2022 Jan 1;50(1):e11-e19. doi: 10.1097/CCM.0000000000005224. PMID: 34582420; PMCID: PMC8678335.

Background

Delirium is an acute and fluctuating disturbance of consciousness, common in hospitalized patients across many medical specialties. Delirium is associated with worse clinical outcomes, including increased length of hospitalization, worse functional outcomes as assessed by the Glasgow Outcome Scale, and increased mortality. Nevertheless, delirium remains largely underdiagnosed. Increasing evidence shows that not only the presence of delirium but also its severity are associated with worse prognosis. Measuring delirium severity is important for assessing prognosis, monitoring response to treatment, and anticipating the burden of care for patients both during and after hospitalization. 

Currently, delirium severity is primarily assessed using clinical behavioral assessments, but these involve intermittent and subjective evaluation of a dynamic, complex condition and can generate disagreement among experts. An automated method that quantifies the presence and severity of delirium directly based on assessment of brain physiology could enable the development of more effective treatments and prevention strategies for delirium.

Early studies showed that qualitative features of electroencephalography (EEG) data are associated with delirium presence and severity. EEG slowing, an increase of delta (1–4 Hz) and/or theta power (4–8 Hz) or a decrease of alpha power (8–12 Hz), correlates with the presence of delirium across various types of delirium presentations. In current practice, EEGs are analyzed using visual interpretation by clinical experts rather than quantitative analysis. Limitations of visual EEG interpretation include interrater variability and the use of only a small number of relatively simple descriptive features, typically scored for their presence or absence. An automated method able to provide a quantitative assessment of the degree of EEG abnormality may provide better monitoring of delirium severity. 

In this work we developed the EEG Confusion Assessment Method Severity (E-CAM-S) score, an automated physiologic method for assessing the presence and severity of delirium using quantitative EEG in a large and heterogeneous patient population. We evaluated which quantitative EEG features are most strongly associated with delirium severity. Last, we investigated whether the E-CAM-S is an independent predictor of important clinical outcomes, including hospital LOS and inhospital mortality.

Methods

Design: Retrospective cohort study.

Setting: Single-center tertiary academic medical center.

Patients: Three-hundred seventy-three adult patients undergoing electroencephalography to evaluate altered mental status between August 2015 and December 2019.

Interventions: None.

Measurements and main results: We developed the E-CAM-S based on a learning-to-rank machine learning model of forehead electroencephalography signals. Clinical delirium severity was assessed using the Confusion Assessment Method Severity (CAM-S). We compared associations of E-CAM-S and CAM-S with hospital length of stay and inhospital mortality. E-CAM-S correlated with clinical CAM-S (R = 0.67; p < 0.0001). For the overall cohort, E-CAM-S and CAM-S were similar in their strength of association with hospital length of stay (correlation = 0.31 vs 0.41, respectively; p = 0.082) and inhospital mortality (area under the curve = 0.77 vs 0.81; p = 0.310). Even when restricted to noncomatose patients, E-CAM-S remained statistically similar to CAM-S in its association with length of stay (correlation = 0.37 vs 0.42, respectively; p = 0.188) and inhospital mortality (area under the curve = 0.83 vs 0.74; p = 0.112). In addition to previously appreciated spectral features, the machine learning framework identified variability in multiple measures over time as important features in electroencephalography-based prediction of delirium severity.

Conclusions: The E-CAM-S is an automated, physiologic measure of delirium severity that predicts clinical outcomes with a level of performance comparable to conventional interview-based clinical assessment.

Data Description

Clinical Data

Patients were assessed for delirium severity at the bedside by study staff. For the patients not undergoing long-term EEG monitoring, it was ensured that the clinical assessment was performed within 1 hour of EEG recording. The delirium severity was assessed using the CAM-S severity scoring method. This tool measures and scores the severity of 4 individual features based on several questions and small assignments: (1) acute change/fluctuating course, 0 to 1 points; (2) inattention, 0 to 2 points; (3) disorganized thinking, 0 to 2 points; and (4) altered level of consciousness, 0 to 2 points. The delirium severity is scored as the sum of the severity of all 4 features (total 0-7 points). Comatose patients represented 33% and were given a CAM-S score of 7. Patients were also evaluated with the Richmond Agitation Sedation Scale (RASS; normal score 0 18) to assess their level of level of consciousness. Clinical outcomes, including length of stay and the in hospital mortality, and Charlson Comorbidity Index scores were extracted and calculated from medical record data.

EEG Recordings

EEGs were recorded with Ag/AgCl scalp electrodes using the standard international 10-20 system for electrode placement, which was performed by qualified EEG technicians. All EEGs had a minimum duration of 20 minutes.

Patient Characteristics

A total of 403 patients were enrolled. Of these patients, 30 were subsequently excluded: 8 due to a prior diagnosis of dementia that was determined after evaluation, 12 due to technical difficulties with the EEG, 4 because of any important missing data and 6 because the time interval between the clinical and EEG test time was too large. 373 patients remained in the dataset, and data from these are provided in this repository.

Of the total of 373 patients analyzed, 251 (67.3%) screened positive for delirium, as defined by CAM-ICU criteria, and 122 (32.7%) of our patients represented comatose patients. When dividing our patient population into delirium and non-delirium patients, it can be seen that patients with delirium in general were more critically ill and had worse clinical outcomes. Patients with delirium in general were older, had lower RASS scores and higher CAM-S and Charlson Comorbidity Index scores. They stayed longer in the hospital and were more likely to experience in-hospital mortality.

Data analysis: See the manuscript: Crit Care Med. 2022 Jan 1;50(1):e11-e19.

Usage Notes

The data are provided in MATLAB (.mat) format.  Each .mat file contains these quantities: 

• Fs - 1x1 - 8 double - sampling frequency              
• channels - 22x4 - 176  char - names of EEG channels
•   data -  22x Nt (Nt is the number of time samples) -  double - the EEG data
•   start_time_shifted - 1x1 - datetime - the (shifted) starting time of the recording

The data can be read using MATLAB or Python. 

Code to read the data and reproduce the results from the manuscript is provided here: 
E-CAM-S Github repository

Ethics

In this dataset, all identifiable patient information has been removed.

Conflicts of Interest

Parts of this work were supported by grants from the NIH (R01NS102190, R01NS102574, R01NS107291, RF1AG064312, RF1NS120947, R01AG073410, R01HL161253, R01NS126282, R01AG073598), and NSF (2014431). M.B.W. is a co-founder of Beacon Biosignals. Beacon Biosignals did not contribute funding nor played any role in the study. 

References

1. van Sleuwen M, Sun H, Eckhardt C, Neelagiri A, Tesh RA, Westmeijer M, Paixao L, Rajan S, Velpula Krishnamurthy P, Sikka P, Leone MJ, Panneerselvam E, Quadri SA, Akeju O, Kimchi EY, Westover MB. Physiological Assessment of Delirium Severity: The Electroencephalographic Confusion Assessment Method Severity Score (E-CAM-S). Crit Care Med. 2022 Jan 1;50(1):e11-e19. doi: 10.1097/CCM.0000000000005224. PMID: 34582420; PMCID: PMC8678335.